MAS: a multiplicative approximation scheme for probabilistic inference.
Ydo WexlerChristopher MeekPublished in: NIPS (2008)
Keyphrases
- probabilistic inference
- graphical models
- bayesian networks
- message passing
- approximate inference
- weighted model counting
- influence diagrams
- belief networks
- multi agent systems
- context specific independence
- conditional probabilities
- mini bucket
- bayesian belief networks
- efficient inference
- variable elimination
- exact inference
- approximation algorithms
- bucket elimination
- machine learning
- np hard
- optimal solution
- belief propagation
- probabilistic model
- special case